Can a robot pass a university entrance exam? | Noriko Arai

202,616 views ใƒป 2017-09-13

TED


์•„๋ž˜ ์˜๋ฌธ์ž๋ง‰์„ ๋”๋ธ”ํด๋ฆญํ•˜์‹œ๋ฉด ์˜์ƒ์ด ์žฌ์ƒ๋ฉ๋‹ˆ๋‹ค.

๋ฒˆ์—ญ: Yousun Kang ๊ฒ€ํ† : JY Kang
00:13
Today, I'm going to talk about AI and us.
0
13014
3660
์˜ค๋Š˜ ์ €๋Š” AI์™€ ์šฐ๋ฆฌ์— ๋Œ€ํ•˜์—ฌ ์ด์•ผ๊ธฐํ•˜๊ณ ์ž ํ•ฉ๋‹ˆ๋‹ค.
00:18
AI researchers have always said
1
18206
2143
AI ์—ฐ๊ตฌ์ž๋“ค์€ ํ•ญ์ƒ
00:20
that we humans do not need to worry,
2
20373
2594
์šฐ๋ฆฌ ์ธ๊ฐ„์€ ๊ฑฑ์ •ํ•  ํ•„์š”๋Š” ์—†๋‹ค๊ณ  ์ฃผ์žฅํ•ด์™”์Šต๋‹ˆ๋‹ค.
00:22
because only menial jobs will be taken over by machines.
3
22991
3580
๋‹จ์ˆœ ๋…ธ๋™์„ ํ•„์š”๋กœ ํ•˜๋Š” ์ง์—…๋งŒ ๊ธฐ๊ณ„๊ฐ€ ๋Œ€์‹ ํ•  ๊ฑฐ๋ผ๋ฉด์„œ ๋ง์ด์ฃ .
00:27
Is that really true?
4
27274
1603
์‚ฌ์‹ค์ผ๊นŒ์š”?
00:30
They have also said that AI will create new jobs,
5
30365
3827
๊ทธ๋ฆฌ๊ณ  AI ๋•๋ถ„์— ์ƒˆ๋กœ์šด ์ง์—…์ด ์ƒ๊ฒจ๋‚˜๊ณ 
00:34
so those who lose their jobs will find a new one.
6
34216
3411
AI ๋•Œ๋ฌธ์— ์ง์—…์„ ์žƒ์€ ์‚ฌ๋žŒ๋„ ์ƒˆ ์ง์—…์„ ์–ป๊ฒŒ ๋  ๊ฑฐ๋ผ๊ณ  ์ฃผ์žฅํ•˜์ฃ .
00:38
Of course.
7
38264
1355
๋งž๋Š” ๋ง์ž…๋‹ˆ๋‹ค.
00:39
But the real question is:
8
39643
2172
ํ•˜์ง€๋งŒ ์—ฌ๊ธฐ์„œ ์งˆ๋ฌธํ•˜๊ณ  ์‹ถ์€ ๊ฒƒ์ด ์žˆ์Šต๋‹ˆ๋‹ค.
00:41
How many of those who may lose their jobs to AI
9
41839
4105
AI ๋•Œ๋ฌธ์— ์ง์—…์„ ์žƒ์€ ์‚ฌ๋žŒ ์ค‘ ๋ช‡ ๋ช…์ด๋‚˜
00:45
will be able to land a new one,
10
45968
2489
์ƒˆ๋กœ์šด ์ผ์— ๋‹ค์‹œ ์ ์‘ํ•  ์ˆ˜ ์žˆ์„๊นŒ์š”?
00:48
especially when AI is smart enough to learn better than most of us?
11
48481
5850
ํŠนํžˆ๋‚˜ AI์˜ ๋ฐœ๋‹ฌ๋กœ ์šฐ๋ฆฌ ์ธ๊ฐ„๋ณด๋‹ค ๋” ๋งŽ์€ ๊ฒƒ์„ ๋ฐฐ์šธ ์ˆ˜ ์žˆ์„ ํ…๋ฐ์š”.
00:55
Let me ask you a question:
12
55397
2185
์—ฌ๋Ÿฌ๋ถ„๊ป˜ ์งˆ๋ฌธ ํ•˜๋‚˜ ํ•˜๊ฒ ์Šต๋‹ˆ๋‹ค.
00:58
How many of you think
13
58666
1798
์—ฌ๋Ÿฌ๋ถ„ ์ค‘์—์„œ
01:00
that AI will pass the entrance examination of a top university by 2020?
14
60488
6094
2020๋…„์ด ๋˜๋ฉด AI๊ฐ€ ์ƒ์œ„๊ถŒ ๋Œ€ํ•™์˜ ์ž…ํ•™ ์‹œํ—˜์— ํ•ฉ๊ฒฉํ•  ์ˆ˜ ์žˆ์œผ๋ฆฌ๋ผ
์ƒ๊ฐํ•˜์‹œ๋Š” ๋ถ„ ๊ณ„์‹ ๊ฐ€์š”?
01:07
Oh, so many. OK.
15
67836
2457
์˜ค, ๋งŽ๊ตฐ์š”. ์ข‹์•„์š”.
01:10
So some of you may say, "Of course, yes!"
16
70317
4358
๋‹น์—ฐํžˆ ํ•ฉ๊ฒฉํ•  ๊ฑฐ๋ผ๊ณ  ๋งŽ์€ ๋ถ„๋“ค์ด ์†์„ ๋“œ์…จ๋Š”๋ฐ์š”.
01:15
Now singularity is the issue.
17
75369
2134
ํ•˜์ง€๋งŒ ๋ชจ๋‘๊ฐ€ ๋˜‘๊ฐ™์ด ์ƒ๊ฐํ•˜๋Š” ๊ฑด ์•„๋‹ˆ์ง€์š”.
01:18
And some others may say, "Maybe,
18
78590
3095
์ด๋ ‡๊ฒŒ ๋งํ•˜๋Š” ๋ถ„๋„ ๊ณ„์‹ค ๊ฑฐ์˜ˆ์š”.
"๊ฐ€๋Šฅํ• ์ง€๋„ ๋ชจ๋ฅด๊ฒ ์–ด. AI๊ฐ€ ์ตœ๊ณ ์˜ ๋ฐ”๋‘‘๊ธฐ์‚ฌ๋„ ์ด๊ฒผ์œผ๋‹ˆ๊นŒ."
01:21
because AI already won against a top Go player."
19
81709
4508
01:27
And others may say, "No, never. Uh-uh."
20
87213
3681
"์ ˆ๋Œ€ ๋ชปํ•  ๊ฑฐ์•ผ!"๋ผ๊ณ  ์ƒ๊ฐํ•˜๋Š” ๋ถ„๋„ ๊ณ„์‹ค ๊ฒ๋‹ˆ๋‹ค.
01:32
That means we do not know the answer yet, right?
21
92195
3593
์ด๋ ‡๊ฒŒ ๋‚˜๋‰˜๋Š” ๊ฑด ์šฐ๋ฆฌ๊ฐ€ ์•„์ง ์ •๋‹ต์„ ๋ชจ๋ฅด๊ธฐ ๋•Œ๋ฌธ์ผ ๊ฑฐ์˜ˆ์š”.
01:36
So that was the reason why I started Todai Robot Project,
22
96268
4960
๊ทธ๋ž˜์„œ ์ €๋Š” ํ† ๋‹ค์ด(Todai) ๋กœ๋ด‡ ํ”„๋กœ์ ํŠธ๋ฅผ ์‹œ์ž‘ํ–ˆ์Šต๋‹ˆ๋‹ค.
01:41
making an AI which passes the entrance examination
23
101252
3872
AI๋ฅผ ์ผ๋ณธ์˜ ์ผ๋ฅ˜๋Œ€ํ•™์ธ ๋„์ฟ„ ๋Œ€ํ•™์˜ ์ž…ํ•™ ์‹œํ—˜์—
01:45
of the University of Tokyo,
24
105148
2589
01:47
the top university in Japan.
25
107761
2537
ํ•ฉ๊ฒฉ์‹œํ‚ค๋ ค๋Š” ํ”„๋กœ์ ํŠธ์ž…๋‹ˆ๋‹ค.
01:51
This is our Todai Robot.
26
111464
2548
์ด๊ฒƒ์ด ํ† ๋‹ค์ด ๋กœ๋ด‡์ž…๋‹ˆ๋‹ค.
01:56
And, of course, the brain of the robot is working in the remote server.
27
116131
5810
๋ฌผ๋ก  ๋กœ๋ด‡์€ ์›๊ฒฉ ์„œ๋ฒ„์— ์—ฐ๊ฒฐ๋˜์–ด ์ž‘๋™ํ•ฉ๋‹ˆ๋‹ค.
02:02
It is now writing a 600-word essay
28
122747
4222
๋ณด๊ณ  ๊ณ„์‹  ๊ฒƒ์€ ์ด ๋กœ๋ด‡์ด
17์„ธ๊ธฐ ํ•ด์ƒ ๋ฌด์—ญ์„ ์ฃผ์ œ๋กœ 600์ž ๋ถ„๋Ÿ‰์˜ ๋…ผ์ˆ ๋ฌธ์„ ์“ฐ๊ณ  ์žˆ๋Š” ๋ชจ์Šต์ž…๋‹ˆ๋‹ค.
02:06
on maritime trade in the 17th century.
29
126993
4119
02:11
How does that sound?
30
131136
1765
์–ด๋– ์‹ ๊ฐ€์š”?
02:14
Why did I take the entrance exam as its benchmark?
31
134113
4104
์™œ ์ž…ํ•™ ์‹œํ—˜์„ ๊ธฐ์ค€์œผ๋กœ ์‚ผ์•˜์„๊นŒ์š”?
02:19
Because I thought we had to study the performance of AI
32
139098
4741
AI์˜ ์„ฑ๊ณผ๋ฅผ ์ธ๊ฐ„๊ณผ ๋น„๊ตํ•˜๊ธฐ ์œ„ํ•ด์„œ๋Š”
02:23
in comparison to humans,
33
143863
2114
์ธ๊ฐ„๋งŒ์ด ํ•  ์ˆ˜ ์žˆ๋‹ค๊ณ  ์—ฌ๊ฒจ์ง€๊ณ 
02:26
especially on the skills and expertise
34
146001
2860
๊ต์œก์„ ํ†ตํ•ด์„œ๋งŒ ํš๋“ํ•  ์ˆ˜ ์žˆ๋‹ค๊ณ  ๋ฏฟ์–ด์กŒ๋˜ ๊ณ ๊ธ‰ ๊ธฐ์ˆ ์„
02:28
which are believed to be acquired only by humans
35
148885
4088
AI๊ฐ€ ์–ด๋–ป๊ฒŒ ์ฒ˜๋ฆฌํ•˜๋Š”์ง€ ์—ฐ๊ตฌํ•ด์•ผ ํ•œ๋‹ค๊ณ 
02:32
and only through education.
36
152997
2335
์ƒ๊ฐํ–ˆ๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค.
02:35
To enter Todai, the University of Tokyo,
37
155782
4043
๋„์ฟ„ ๋Œ€ํ•™์— ์ž…ํ•™ํ•˜๊ธฐ ์œ„ํ•ด์„œ๋Š”
02:39
you have to pass two different types of exams.
38
159849
4421
๋‘ ๋ฒˆ์˜ ์‹œํ—˜์„ ํ†ต๊ณผํ•ด์•ผ ํ•ฉ๋‹ˆ๋‹ค.
02:44
The first one is a national standardized test
39
164294
3760
์ฒซ์งธ๋กœ๋Š” ๊ฐ๊ด€์‹์œผ๋กœ ๋˜์–ด์žˆ๋Š”
02:48
in multiple-choice style.
40
168078
2403
๊ตญ๊ฐ€ ๊ณต์ธ ์‹œํ—˜์„ ํ’€์–ด์•ผ ํ•ฉ๋‹ˆ๋‹ค.
02:50
You have to take seven subjects
41
170505
2455
7 ๊ณผ๋ชฉ์˜ ์‹œํ—˜์„ ์น˜๋ฃจ๋Š”๋ฐ
02:52
and achieve a high score --
42
172984
1955
๋†’์€ ์„ฑ์ ์„ ๋ฐ›์•„์•ผ ํ•˜์ง€์š”.
02:54
I would say like an 85 percent or more accuracy rate --
43
174963
4772
84์  ์ด์ƒ์€ ๋˜์–ด์•ผ
02:59
to be allowed to take the second stage written test
44
179759
4087
๋‹ค์Œ ๋‹จ๊ณ„์ธ ์ž‘๋ฌธ ์‹œํ—˜์„ ๋ณผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
03:03
prepared by Todai.
45
183870
2159
๋„์ฟ„๋Œ€ํ•™์ด ์ง์ ‘ ์ถœ์ œํ•˜๋Š” ์‹œํ—˜์ด์ฃ .
03:06
So let me first explain how modern AI works,
46
186994
5317
ํ˜„๋Œ€์˜ AI๊ฐ€ ์–ด๋–ป๊ฒŒ ์ž‘์—…ํ•ด์„œ "๋‹จ์–ด ์—ฐ์ƒ ํ€ด์ฆˆ"๋ฅผ ํ‘ธ๋Š”์ง€
03:12
taking the "Jeopardy!" challenge as an example.
47
192335
3069
์˜ˆ๋ฅผ ๋“ค์–ด ์„ค๋ช…ํ•˜๊ฒ ์Šต๋‹ˆ๋‹ค.
03:17
Here is a typical "Jeopardy!" question:
48
197539
3079
"๋‹จ์–ด ์—ฐ์ƒ ํ€ด์ฆˆ"์˜ ์ „ํ˜•์ ์ธ ๋ฌธ์ œ ํ˜•ํƒœ๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™์Šต๋‹ˆ๋‹ค.
03:20
"Mozart's last symphony shares its name with this planet."
49
200642
4461
"๋ชจ์งœ๋ฅดํŠธ์˜ ๋งˆ์ง€๋ง‰ ๊ตํ–ฅ๊ณก ์ œ๋ชฉ์€ ์ด ํ–‰์„ฑ์˜ ์ด๋ฆ„๊ณผ ๊ฐ™๋‹ค."
03:26
Interestingly, a "Jeopardy!" question always asks,
50
206195
4013
ํฅ๋ฏธ๋กญ๊ฒŒ๋„ "๋‹จ์–ด ์—ฐ์ƒ ํ€ด์ฆˆ" ๋ฌธ์ œ๋Š” ํ•ญ์ƒ "์ด๊ฒƒ"์œผ๋กœ ์งˆ๋ฌธํ•˜๊ณ 
03:30
always ends with "this" something:
51
210232
3328
"์ด๊ฒƒ"์œผ๋กœ ๋๋‚ฉ๋‹ˆ๋‹ค.
03:33
"this" planet, "this" country,
52
213584
2827
"์ด" ํ–‰์„ฑ, "์ด" ๊ตญ๊ฐ€,
03:36
"this" rock musician, and so on.
53
216435
2608
"์ด" ๋ก ๋ฎค์ง€์…˜ ๋“ฑ๋“ฑ์ด์ง€์š”.
03:39
In other words, "Jeopardy!" doesn't ask many different types of questions,
54
219067
4299
๋‹ค์‹œ ๋งํ•˜์ž๋ฉด, "๋‹จ์–ด ์—ฐ์ƒ ํ€ด์ฆˆ"๋Š” ์—ฌ๋Ÿฌ๊ฐ€์ง€ ๋‹ค๋ฅธ ํ˜•ํƒœ์˜ ์งˆ๋ฌธ์ด ์•„๋‹ˆ๋ผ
03:43
but a single type,
55
223390
1837
ํ•˜๋‚˜์˜ ํ˜•ํƒœ๋กœ
03:45
which we call "factoid questions."
56
225251
2536
"ํŒฉํ† ์ด๋“œ(Factoid) ๋ฌธ์ œ"๋ผ๊ณ  ํ•ฉ๋‹ˆ๋‹ค.
03:48
By the way, do you know the answer?
57
228975
2167
์•„๊นŒ ๊ทธ ์งˆ๋ฌธ์˜ ์ •๋‹ต์„ ์•„์‹œ๋‚˜์š”?
03:53
If you do not know the answer and if you want to know the answer,
58
233980
4055
๋งŒ์•ฝ ์ •๋‹ต์„ ๋ชจ๋ฅด๊ฑฐ๋‚˜ ๋‹ต์„ ์•Œ๊ณ  ์‹ถ์„ ๋•Œ๋Š”
03:58
what would you do?
59
238059
1287
์—ฌ๋Ÿฌ๋ถ„์€ ์–ด๋–ป๊ฒŒ ํ•˜์‹œ๋‚˜์š”?
04:00
You Google, right? Of course.
60
240160
3132
๊ฒ€์ƒ‰์„ ํ•˜์ง€์š”.
04:03
Why not?
61
243316
1480
๋‹น์—ฐํ•˜์ฃ .
04:04
But you have to pick appropriate keywords
62
244820
3592
ํ•˜์ง€๋งŒ ์ ์ ˆํ•œ ๊ฒ€์ƒ‰์–ด๋ฅผ ์ž…๋ ฅํ•ด์•ผ ํ•ฉ๋‹ˆ๋‹ค.
04:08
like "Mozart," "last" and "symphony" to search.
63
248436
4364
"๋ชจ์งœ๋ฅดํŠธ"๋‚˜ "๋งˆ์ง€๋ง‰", "๊ตํ–ฅ๊ณก" ๊ฐ™์€ ๊ฒ€์ƒ‰์–ด๊ฐ€ ํ•„์š”ํ•˜์ฃ .
04:13
The machine basically does the same.
64
253462
2400
๋กœ๋ด‡์˜ ๊ฒฝ์šฐ๋„ ๋น„์Šทํ•ฉ๋‹ˆ๋‹ค.
04:16
Then this Wikipedia page will be ranked top.
65
256457
4660
์œ„ํ‚คํ”ผ๋””์•„ ์‚ฌ์ดํŠธ๊ฐ€ ์ œ์ผ ๋จผ์ € ๊ฒ€์ƒ‰๋  ๊ฒ๋‹ˆ๋‹ค.
04:21
Then the machine reads the page.
66
261840
1908
๊ทธ๋ฆฌ๊ณ  ๊ทธ ํŽ˜์ด์ง€๋ฅผ ์ฝ์–ด๋ณด๋ฉด ๋˜๊ฒ ์ฃ .
04:23
No, uh-uh.
67
263772
1171
์‚ฌ์‹ค์€ ์•„๋‹™๋‹ˆ๋‹ค.
04:25
Unfortunately, none of the modern AIs,
68
265470
3462
๋ถˆํ–‰ํ•˜๊ฒŒ๋„ ํ˜„๋Œ€์˜ AI ์ค‘์—์„œ
04:28
including Watson, Siri and Todai Robot,
69
268956
3968
์™“์Šจ, ์‹œ๋ฆฌ, ํ† ๋‹ค์ด ๋กœ๋ด‡ ๋“ฑ ๊ทธ ์–ด๋–ค ๊ฒƒ๋„
04:32
is able to read.
70
272948
1661
์•„์ง ์ฝ๋Š” ๋Šฅ๋ ฅ์€ ์—†์Šต๋‹ˆ๋‹ค.
04:35
But they are very good at searching and optimizing.
71
275437
3800
ํ•˜์ง€๋งŒ ๊ฒ€์ƒ‰ํ•˜๊ณ  ๊ฒฐ๊ณผ๋ฅผ ์ตœ์ ํ™”ํ•˜๋Š” ๋Šฅ๋ ฅ์€ ๋›ฐ์–ด๋‚˜์ฃ .
04:40
It will recognize
72
280158
2023
๋”ฐ๋ผ์„œ ๋กœ๋ด‡์€
04:42
that the keywords "Mozart," "last" and "symphony"
73
282866
2935
"๋ชจ์งœ๋ฅดํŠธ", "๋งˆ์ง€๋ง‰", "๊ตํ–ฅ๊ณก" ๊ฐ™์€ ๊ฒ€์ƒ‰์–ด๊ฐ€
04:45
are appearing heavily around here.
74
285825
2903
์ฃผ๋กœ ๋‚˜ํƒ€๋‚œ๋‹ค๋Š” ์‚ฌ์‹ค์„ ์•Œ์•„๋ƒ…๋‹ˆ๋‹ค.
04:49
So if it can find a word which is a planet
75
289790
4375
๋”ฐ๋ผ์„œ ๋งŒ์•ฝ ํ–‰์„ฑ ์ด๋ฆ„์„ ๋ฐœ๊ฒฌํ•˜๊ณ 
04:54
and which is co-occurring with these keywords,
76
294189
3648
๊ทธ ํ–‰์„ฑ์ด ์ด ๊ฒ€์ƒ‰์–ด๋“ค๊ณผ ํ•จ๊ป˜ ๊ฒ€์ƒ‰๋œ๋‹ค๋ฉด
04:57
that must be the answer.
77
297861
1989
์ •๋‹ต์ด๋ผ๊ณ  ์ƒ๊ฐํ•˜์ง€์š”.
05:00
This is how Watson finds the answer "Jupiter," in this case.
78
300762
5186
์™“์Šจ์€ ์ •๋‹ต์ธ "๋ชฉ์„ฑ"์„ ์ด๋Ÿฐ ๋ฐฉ๋ฒ•์œผ๋กœ ์ฐพ์•„๋ƒ…๋‹ˆ๋‹ค.
05:08
Our Todai Robot works similarly, but a bit smarter
79
308433
4049
ํ† ๋‹ค์ด ๋กœ๋ด‡๋„ ๋น„์Šทํ•˜๊ฒŒ ์ž‘๋™ํ•˜์ง€๋งŒ
์—ญ์‚ฌ์— ๊ด€ํ•œ ์˜ˆ/์•„๋‹ˆ์˜ค ๋ฌธ์ œ๋ฅผ ํ’€ ๋•Œ ์ข€ ๋” ๋˜‘๋˜‘ํ•œ ๋ฐฉ์‹์œผ๋กœ ์ฒ˜๋ฆฌํ•ฉ๋‹ˆ๋‹ค.
05:12
in answering history yes-no questions,
80
312506
3239
05:16
like, "'Charlemagne repelled the Magyars.' Is this sentence true or false?"
81
316560
5663
"์ƒค๋ฅผ๋งˆ๋‰ด๋Š” ๋งˆ์ž๋ฅด์ธ์„ ์ซ’์•„๋ƒˆ๋‹ค" ์ด ๋ฌธ์ œ๋Š” ์‚ฌ์‹ค์ผ๊นŒ์š”, ๊ฑฐ์ง“์ผ๊นŒ์š”?
05:23
Our robot starts producing a factoid question,
82
323181
4073
์ด ๋ฌธ์žฅ์„ ๋ณด๊ณ  ํ† ๋‹ค์ด ๋กœ๋ด‡์€ ํŒฉํ† ์ด๋“œ ๋ฌธ์ œ๋กœ ์žฌ๊ตฌ์„ฑํ•ฉ๋‹ˆ๋‹ค.
05:27
like: "Charlemagne repelled [this person type]" by itself.
83
327278
4899
๋กœ๋ด‡์€ "์ƒค๋ฅผ๋งˆ๋‰ด๋Š” [๋ˆ„๊ตฌ๋ˆ„๊ตฌ๋ฅผ] ์ซ’์•„๋ƒˆ๋‹ค."๋ผ๊ณ  ์Šค์Šค๋กœ ๋งŒ๋“ค์–ด๋‚ด์ง€์š”.
05:32
Then, "Avars" but not "Magyars" is ranked top.
84
332995
4732
๊ทธ๋Ÿฌ๋ฉด "๋งˆ์ž๋ฅด์ธ"์ด ์•„๋‹ˆ๋ผ "์•„๋ฐ”๋ฅด์ธ"์ด ๊ฒ€์ƒ‰๋ฉ๋‹ˆ๋‹ค.
05:38
This sentence is likely to be false.
85
338357
3049
๊ฒฐ๊ตญ, ์ด ๋ฌธ์žฅ์€ ๊ฑฐ์ง“์ด๋ผ๊ณ  ํŒ๋‹จํ•˜๊ฒŒ ๋˜์ง€์š”.
05:42
Our robot does not read, does not understand,
86
342772
4860
์šฐ๋ฆฌ ๋กœ๋ด‡์€ ์ฝ์ง€๋„, ์ดํ•ดํ•˜์ง€๋„ ๋ชปํ•˜์ง€๋งŒ
05:48
but it is statistically correct in many cases.
87
348335
4144
ํ†ต๊ณ„์ ์œผ๋กœ ๋ณด๋ฉด ๋Œ€๋ถ€๋ถ„ ์ •๋‹ต์„ ๋งž์ถฅ๋‹ˆ๋‹ค.
05:54
For the second stage written test,
88
354147
2508
๋‘˜์งธ ๋‹จ๊ณ„์ธ ์ž‘๋ฌธ ์‹œํ—˜์—์„œ๋Š”
05:56
it is required to write a 600-word essay like this one:
89
356679
5106
๋‹ค์Œ๊ณผ ๊ฐ™์ด 600์ž ๋…ผ์ˆ ๋ฌธ์„ ์จ์•ผ ํ•ฉ๋‹ˆ๋‹ค.
06:01
[Discuss the rise and fall of the maritime trade
90
361809
2278
[17์„ธ๊ธฐ ๋™์•„์‹œ์•„์™€ ๋™๋‚จ์•„์‹œ์•„ ๊ฐ„์˜
06:04
in East and Southeast Asia in the 17th century ...]
91
364111
2422
ํ•ด์ƒ ๋ฌด์—ญ์˜ ํฅ๋ง๊ณผ ์„ฑ์‡ ๋ฅผ ๋…ผํ•˜์‹œ์˜ค.]
06:06
and as I have shown earlier,
92
366557
1387
๊ทธ๋Ÿฌ๋ฉด ์•ž์—์„œ ๋ง์”€๋“œ๋ฆฐ ๊ฒƒ๊ณผ ๋งˆ์ฐฌ๊ฐ€์ง€๋กœ
06:07
our robot took the sentences from the textbooks and Wikipedia,
93
367968
4194
๋กœ๋ด‡์€ ์œ„ํ‚คํ”ผ๋””์•„์™€ ๊ต๊ณผ์„œ์˜ ๋ฌธ์žฅ์„ ๊ฒ€์ƒ‰ํ•˜๊ณ 
06:12
combined them together,
94
372186
1961
๊ฒฐํ•ฉํ•˜๊ณ  ์ตœ์ ํ™”ํ•ด์„œ
06:14
and optimized it to produce an essay
95
374171
3619
ํ•˜๋‚˜์˜ ๋…ผ์ˆ ๋ฌธ์„ ์™„์„ฑํ•ฉ๋‹ˆ๋‹ค.
06:17
without understanding a thing.
96
377814
2207
๋‚ด์šฉ์€ ํ•˜๋‚˜๋„ ์ดํ•ดํ•˜์ง€ ๋ชปํ•˜๋ฉด์„œ ๋ง์ด์ฃ .
06:20
(Laughter)
97
380045
1737
(์›ƒ์Œ)
06:21
But surprisingly, it wrote a better essay
98
381806
4895
ํ•˜์ง€๋งŒ ๋†€๋ผ์šด ๊ฒƒ์€
๋Œ€๋ถ€๋ถ„์˜ ํ•™์ƒ๋“ค์ด ์“ด ๋…ผ์ˆ ๋ฌธ๋ณด๋‹ค ๋‚ซ๋‹ค๋Š” ์‚ฌ์‹ค์ž…๋‹ˆ๋‹ค.
06:26
than most of the students.
99
386725
1561
06:28
(Laughter)
100
388310
2391
(์›ƒ์Œ)
06:30
How about mathematics?
101
390725
1529
์ˆ˜ํ•™์€ ์–ด๋–จ๊นŒ์š”?
06:33
A fully automatic math-solving machine
102
393354
3158
"์ธ๊ณต ์ง€๋Šฅ"์ด๋ผ๋Š” ๋‹จ์–ด๊ฐ€ ํƒ„์ƒํ•œ ํ›„๋กœ
06:36
has been a dream
103
396536
1631
์ „์ž๋™์œผ๋กœ ์ˆ˜ํ•™ ๋ฌธ์ œ๋ฅผ ํ’€ ์ˆ˜ ์žˆ๋Š” ๊ธฐ๊ณ„๋ฅผ ๋งŒ๋“ค ์ˆ˜ ์žˆ์„ ๊ฑฐ๋ผ๊ณ 
06:38
since the birth of the word "artificial intelligence,"
104
398191
4679
๊ฟˆ๊ฟ”์™”์Šต๋‹ˆ๋‹ค.
06:43
but it has stayed at the level of arithmetic for a long, long time.
105
403785
6007
ํ•˜์ง€๋งŒ ๊ณ„์‚ฐํ•  ์ˆ˜ ์žˆ๋Š” ์—ฐ์‚ฐ์˜ ์ˆ˜์ค€์€ ์˜ค๋žซ๋™์•ˆ ๋”์ด์ƒ ๋ฐœ์ „ํ•˜์ง€ ๋ชปํ–ˆ์ง€์š”.
06:51
Last year, we finally succeeded in developing a system
106
411530
5350
์ง€๋‚œํ•ด์—๋Š” ์šฐ๋ฆฌ๊ฐ€ ๊ฐœ๋ฐœํ•œ ์‹œ์Šคํ…œ์ด
06:56
which solved pre-university-level problems from end to end,
107
416904
5173
์˜ˆ๋น„ ๋Œ€ํ•™ ์ˆ˜์ค€์˜ ๋ฌธ์ œ๋ฅผ ์ฒ˜์Œ๋ถ€ํ„ฐ ๋๊นŒ์ง€ ํ‘ธ๋Š” ๋ฐ์— ๋งˆ์นจ๋‚ด ์„ฑ๊ณตํ–ˆ์Šต๋‹ˆ๋‹ค.
07:02
like this one.
108
422101
1262
๋ฐ”๋กœ ์ด๋Ÿฐ ๋ฌธ์ œ์ฃ .
07:05
This is the original problem written in Japanese,
109
425648
4002
์ด๊ฒƒ์€ ์ผ๋ณธ์–ด๋กœ ๋œ ์›๋ž˜์˜ ๋ฌธ์ œ์ž…๋‹ˆ๋‹ค.
07:09
and we had to teach it 2,000 mathematical axioms
110
429674
4397
์šฐ๋ฆฌ๋Š” 2,000๊ฐ€์ง€์˜ ์ˆ˜ํ•™์  ์›๋ฆฌ์™€
07:14
and 8,000 Japanese words
111
434095
2774
8,000๊ฐœ์˜ ์ผ๋ณธ์–ด ๋‹จ์–ด๋ฅผ ๊ฐ€๋ฅด์ณ์•ผ ํ–ˆ์Šต๋‹ˆ๋‹ค.
07:16
to make it accept the problems written in natural language.
112
436893
4558
๊ทธ๋ž˜์•ผ ๊ธฐ๊ณ„๊ฐ€ ์–ธ์–ด๋ฅผ ๋ฐ›์•„๋“ค์ด๊ณ  ๋ฌธ์ œ๋ฅผ ์ˆ˜์šฉํ•  ์ˆ˜ ์žˆ์œผ๋‹ˆ๊นŒ์š”.
07:22
And it is now translating the original problems
113
442234
3542
์ด์ œ ๋ฌธ์ œ๋ฅผ ๊ธฐ๊ณ„๊ฐ€ ํ•ด๋…ํ•  ์ˆ˜ ์žˆ๋Š” ํ˜•ํƒœ๋กœ
07:25
into machine-readable formulas.
114
445800
3139
๋ฒˆ์—ญ์„ ํ•ฉ๋‹ˆ๋‹ค.
07:30
Weird, but it is now ready to solve it, I think.
115
450578
6099
์ด์ƒํ•˜๊ฒŒ ๋ณด์ด๊ธด ํ•˜์ง€๋งŒ, ์ด์ œ ํ’€ ์ค€๋น„๊ฐ€ ๋œ ๊ฒƒ ๊ฐ™๊ตฐ์š”.
07:36
Go and solve it.
116
456701
1411
์ž, ๋ฌธ์ œ๋ฅผ ํ’‰๋‹ˆ๋‹ค.
07:38
Yes! It is now executing symbolic computation.
117
458818
4284
์ด์ œ ๊ธฐํ˜ธํ™” ๋œ ๊ณ„์‚ฐ์„ ์‹คํ–‰ํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.
07:44
Even more weird,
118
464030
1580
๋” ์ด์ƒํ•˜๊ฒŒ ๋ณด์ผ ๊ฒ๋‹ˆ๋‹ค.
07:45
but probably this is the most fun part for the machine.
119
465634
4825
๊ฐ€์žฅ ์žฌ๋ฏธ์žˆ๋Š” ๋ถ€๋ถ„์ธ๋ฐ์š”.
07:50
(Laughter)
120
470483
2351
(์›ƒ์Œ)
07:52
Now it outputs a perfect answer,
121
472858
2815
๋“œ๋””์–ด ์ •๋‹ต์„ ๋งž์ถ”์—ˆ์Šต๋‹ˆ๋‹ค.
07:55
though its proof is impossible to read, even for mathematicians.
122
475697
4707
์ˆ˜ํ•™ ์ „๊ณต์ž๋„ ์ดํ•ด ๋ชปํ•˜๋Š” ์ฆ๋ช…์‹์œผ๋กœ ๋ง์ด์ง€์š”.
08:02
Anyway, last year our robot was among the top one percent
123
482773
6961
์–ด์จŒ๋“  ์ž‘๋…„์— ์šฐ๋ฆฌ ๋กœ๋ด‡์€ 2๋‹จ๊ณ„ ์ˆ˜ํ•™ ๋…ผ์ˆ  ์‹œํ—˜์—์„œ
08:10
in the second stage written exam in mathematics.
124
490199
3633
์ƒ์œ„ 1%์— ๋“ค์—ˆ์Šต๋‹ˆ๋‹ค.
08:14
(Applause)
125
494652
3210
(๋ฐ•์ˆ˜)
08:18
Thank you.
126
498412
1311
๊ฐ์‚ฌํ•ฉ๋‹ˆ๋‹ค.
08:19
So, did it enter Todai?
127
499747
2471
๊ทธ๋ž˜์„œ ํ† ๋‹ค์ด๋Š” ๋Œ€ํ•™์— ์ž…ํ•™์„ ํ•  ์ˆ˜ ์žˆ์—ˆ์„๊นŒ์š”?
08:22
No, not as I expected.
128
502981
3058
์˜ˆ์ƒ๊ณผ ๋‹ฌ๋ฆฌ ๋ชปํ–ˆ์Šต๋‹ˆ๋‹ค.
08:26
Why?
129
506783
1399
์ด์œ ๋Š” ๋ฌด์—‡์ผ๊นŒ์š”?
08:28
Because it doesn't understand any meaning.
130
508206
2639
์˜๋ฏธ๋ฅผ ์ „ํ˜€ ์ดํ•ดํ•˜์ง€ ๋ชปํ–ˆ๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค.
08:32
Let me show you a typical error it made in the English test.
131
512308
4079
์˜์–ด ์‹œํ—˜์—์„œ ๋กœ๋ด‡์€ ๋‹ค์Œ๊ณผ ๊ฐ™์€ ์‹ค์ˆ˜๋ฅผ ์ €์ง€๋ฆ…๋‹ˆ๋‹ค.
08:36
[Nate: We're almost at the bookstore. Just a few more minutes.
132
516411
2977
[๋„ค์ดํŠธ: ์„œ์ ์— ๊ฑฐ์˜ ๋„์ฐฉํ–ˆ์–ด. ๋ช‡ ๋ถ„๋งŒ ๋” ๊ฐ€๋ฉด ๋ผ.]
08:39
Sunil: Wait. ______ . Nate: Thank you! That always happens ...]
133
519412
3039
[์„ ์ผ: ์ž ๊น. _________. ๋„ค์ดํŠธ: ๊ณ ๋งˆ์›Œ. ๋‚œ ํ•ญ์ƒ ์ด๋Ÿฐ๋‹ค....]
08:42
Two people are talking.
134
522475
1151
๋‘ ์‚ฌ๋žŒ์ด ๋Œ€ํ™”๋ฅผ ํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.
08:43
For us, who can understand the situation --
135
523650
2054
์ด ์ƒํ™ฉ์„ ์ดํ•ดํ–ˆ๋‹ค๋ฉด ์ •๋‹ต์€..
08:45
[1. "We walked for a long time." 2. "We're almost there."
136
525704
2773
[1. "์šฐ๋ฆฌ ์˜ค๋žซ๋™์•ˆ ๊ฑธ์—ˆ์–ด." 2. "๊ฑฐ์˜ ๋‹ค์™”๋‹ค."
08:48
3. "Your shoes look expensive." 4. "Your shoelace is untied."]
137
528501
3032
3. "๋„ˆ ์‹ ๋ฐœ ๋น„์‹ธ๋ณด์ธ๋‹ค." 4. "๋„ˆ ์‹ ๋ฐœ ๋ˆ ํ’€๋ ธ์–ด."]
08:51
it is obvious number four is the correct answer, right?
138
531557
2873
์ •๋‹ต์€ 4๋ฒˆ์ด์ง€์š”.
08:54
But Todai Robot chose number two,
139
534454
2238
ํ•˜์ง€๋งŒ ํ† ๋‹ค์ด ๋กœ๋ด‡์€ 2๋ฒˆ์ด๋ผ๊ณ  ๋‹ตํ–ˆ์Šต๋‹ˆ๋‹ค.
08:56
even after learning 15 billion English sentences
140
536716
5360
๋”ฅ๋Ÿฌ๋‹(deep learning) ๊ธฐ์ˆ ๋กœ 150์–ต ๊ฐœ์˜ ์˜์–ด ๋ฌธ์žฅ์„
09:02
using deep learning technologies.
141
542100
2728
๋ฐฐ์› ์Œ์—๋„ ๋ถˆ๊ตฌํ•˜๊ณ  ๋ง์ด์ฃ .
09:07
OK, so now you might understand what I said:
142
547600
4172
์—ฌ๋Ÿฌ๋ถ„๋„ ์ด์ œ ์ œ ์–˜๊ธฐ๊ฐ€ ์ดํ•ด๋˜์‹ค ๊ฑฐ์˜ˆ์š”.
09:12
modern AIs do not read,
143
552399
2648
ํ˜„๋Œ€์˜ AI๋Š” ์ฝ์ง€๋„
09:15
do not understand.
144
555071
1413
์ดํ•ดํ•˜์ง€๋„ ๋ชปํ•œ๋‹ค๋Š” ๊ฒƒ์„์š”.
09:17
They only disguise as if they do.
145
557516
3169
์ฝ๊ณ  ์ดํ•ดํ•˜๋Š” ๊ฒƒ์ฒ˜๋Ÿผ ํ–‰๋™ํ•  ๋ฟ์ž…๋‹ˆ๋‹ค.
09:24
This is the distribution graph
146
564867
2981
์ด๊ฒƒ์€ ํ† ๋‹ค์ด ๋กœ๋ด‡๊ณผ ๊ฐ™์€ ์‹œํ—˜์„ ์น˜๋ฃฌ
09:27
of half a million students who took the same exam as Todai Robot.
147
567872
5777
50๋งŒ๋ช… ํ•™์ƒ๋“ค์˜ ๋ถ„ํฌ๋„์ž…๋‹ˆ๋‹ค.
09:34
Now our Todai Robot is among the top 20 percent,
148
574558
5165
์šฐ๋ฆฌ ํ† ๋‹ค์ด ๋กœ๋ด‡์ด ์ƒ์œ„ 20% ๋‚ด์— ์žˆ๊ธฐ ๋•Œ๋ฌธ์—
09:40
and it was capable to pass
149
580986
2415
์ผ๋ณธ์— ์žˆ๋Š” ๋Œ€ํ•™ ์ค‘์—์„œ
09:43
more than 60 percent of the universities in Japan --
150
583425
3941
60% ์ด์ƒ์˜ ๋Œ€ํ•™์— ์ž…ํ•™ํ•  ์ˆ˜ ์žˆ์—ˆ์ง€๋งŒ
09:47
but not Todai.
151
587390
1377
๋™๊ฒฝ๋Œ€ ์ž…ํ•™์€ ์‹คํŒจํ–ˆ์Šต๋‹ˆ๋‹ค.
09:50
But see how it is beyond the volume zone
152
590116
4025
ํ•˜์ง€๋งŒ ๋ฏธ๋ž˜์— ์‚ฌ๋ฌด์ง์— ์ข…์‚ฌํ•˜๊ฒŒ ๋ 
09:54
of to-be white-collar workers.
153
594165
2864
์ค‘์‚ฐ์ธต ๊ณ„๊ธ‰ ์ด์ƒ์˜ ์ˆ˜์ค€์ด์ง€์š”.
10:00
You might think I was delighted.
154
600060
2858
์ œ๊ฐ€ ์ด ์„ฑ๊ณผ๋กœ ๊ธฐ๋ปํ–ˆ์„๊นŒ์š”?
10:03
After all, my robot was surpassing students everywhere.
155
603939
3971
์ œ ๋กœ๋ด‡์ด ์—ฌ๋Š ํ•™์ƒ๋“ค๋ณด๋‹ค ๋›ฐ์–ด๋‚  ํ…Œ๋‹ˆ๊นŒ์š”.
10:09
Instead, I was alarmed.
156
609022
2691
๊ทธ์™€ ๋ฐ˜๋Œ€๋กœ ์ „ ๋‘๋ ค์› ์Šต๋‹ˆ๋‹ค.
10:13
How on earth could this unintelligent machine outperform students --
157
613086
5607
์ด๋ ‡๊ฒŒ ๋ฌด์‹ํ•œ ๊ธฐ๊ณ„๊ฐ€ ํ•™์ƒ๋“ค์„ ์ด๊ธด๋‹ค๋Š” ๊ฒŒ ๋ง์ด ๋˜๋‚˜์š”?
10:18
our children?
158
618717
1292
์šฐ๋ฆฌ ์•„์ด๋“ค์„์š”.
10:20
Right?
159
620033
1153
๊ทธ๋ ‡์ง€ ์•Š๋‚˜์š”?
10:22
I decided to investigate what was going on in the human world.
160
622101
4402
์ €๋Š” ์ธ๊ฐ„ ์„ธ์ƒ์— ๋ฌด์Šจ ์ผ์ด ์ผ์–ด๋‚˜๊ณ  ์žˆ๋Š”์ง€ ์กฐ์‚ฌํ•˜๊ธฐ๋กœ ํ–ˆ์Šต๋‹ˆ๋‹ค.
10:28
I took hundreds of sentences from high school textbooks
161
628542
4729
๊ณ ๋“ฑํ•™๊ต ๊ต๊ณผ์„œ์—์„œ ์ˆ˜๋ฐฑ ๊ฐœ์˜ ๋ฌธ์žฅ์„ ๋ฝ‘์•„
10:33
and made easy multiple-choice quizzes,
162
633859
3313
์‰ฌ์šด ๊ฐ๊ด€์‹ ๋ฌธ์ œ๋ฅผ ๋งŒ๋“ค์—ˆ๊ณ 
10:37
and asked thousands of high school students to answer.
163
637196
4143
์ˆ˜์ฒœ ๋ช…์˜ ๊ณ ๋“ฑํ•™์ƒ๋“ค์—๊ฒŒ ๋‹ต์„ ๊ณ ๋ฅด๋„๋ก ํ•ด๋ดค์Šต๋‹ˆ๋‹ค.
10:42
Here is an example:
164
642690
1176
์˜ˆ๋ฅผ ๋“ค๋ฉด ์ด๋Ÿฐ ๊ฒƒ์ž…๋‹ˆ๋‹ค.
10:43
[Buddhism spread to ... , Christianity to ... and Oceania,
165
643890
2818
[๋ถˆ๊ต๋Š” ๋™๋‚จ์•„์‹œ์•„์™€ ๋™์•„์‹œ์•„๋กœ ํผ์กŒ๋‹ค.
๊ธฐ๋…๊ต๋Š” ์œ ๋Ÿฝ, ๋ถ๋ฏธ, ๋‚จ๋ฏธ, ์˜ค์„ธ์•„๋‹ˆ์•„๋กœ ํผ์กŒ๋‹ค. ์ด์Šฌ๋žŒ์€- ]
10:46
and Islam to ...]
166
646732
1151
10:47
Of course, the original problems are written in Japanese,
167
647907
2740
๋ฌผ๋ก  ๋ฌธ์ œ๋Š” ์ผ๋ณธ์–ด ํ˜น์€ ํ•™์ƒ๋“ค์˜ ๋ชจ๊ตญ์–ด์˜€๊ณ ์š”.
10:50
their mother tongue.
168
650671
1155
10:51
[ ______ has spread to Oceania.
169
651850
1515
[______์€ ์˜ค์„ธ์•„๋‹ˆ์•„๋กœ ํผ์กŒ๋‹ค.
10:53
1. Hinduism 2. Christianity 3. Islam 4. Buddhism ]
170
653389
2417
1. ํžŒ๋‘๊ต 2. ๊ธฐ๋…๊ต 3. ์ด์Šฌ๋žŒ๊ต 4. ๋ถˆ๊ต]
10:55
Obviously, Christianity is the answer, isn't it?
171
655830
2299
์ •๋‹ต์€ ๋ฌผ๋ก  ๊ธฐ๋…๊ต์ž…๋‹ˆ๋‹ค.
10:58
It's written!
172
658153
1214
๋ฌธ์ œ์— ์ด๋ฏธ ๋‚˜์™€ ์žˆ์ฃ .
11:01
And Todai Robot chose the correct answer, too.
173
661482
4026
ํ† ๋‹ค์ด ๋กœ๋ด‡ ์—ญ์‹œ ์ •๋‹ต์„ ๋งž์ถ”์—ˆ์Šต๋‹ˆ๋‹ค.
11:06
But one-third of junior high school students
174
666758
4879
ํ•˜์ง€๋งŒ 3๋ถ„์˜ 1์˜ ์ค‘ํ•™์ƒ๋“ค์€
11:11
failed to answer this question.
175
671661
2612
์˜ค๋‹ต์„ ์„ ํƒํ–ˆ์Šต๋‹ˆ๋‹ค.
11:16
Do you think it is only the case in Japan?
176
676456
3159
์ผ๋ณธ๋งŒ์˜ ๋ฌธ์ œ์ผ๊นŒ์š”?
11:19
I do not think so,
177
679639
1976
์•„๋‹ˆ๋ผ๊ณ  ์ƒ๊ฐํ•ฉ๋‹ˆ๋‹ค.
11:21
because Japan is always ranked among the top in OECD PISA tests,
178
681639
6371
15์„ธ ํ•™์ƒ๋“ค์˜ ์ˆ˜ํ•™, ๊ณผํ•™, ์ฝ๊ธฐ ์‹ค๋ ฅ์„ ์ธก์ •ํ•˜๊ธฐ ์œ„ํ•ด
3๋…„๋งˆ๋‹ค ์—ด๋ฆฌ๋Š” OECD PISA ์‹œํ—˜์—์„œ
11:28
measuring 15-year-old students' performance in mathematics,
179
688034
3927
์ผ๋ณธ ํ•™์ƒ๋“ค์€ ํ•ญ์ƒ ์ƒ์œ„๊ถŒ์„ ์œ ์ง€ํ•˜๊ณ  ์žˆ๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค.
11:31
science and reading
180
691985
1964
11:33
every three years.
181
693973
1636
11:39
We have been believing
182
699390
2053
์šฐ๋ฆฌ๋Š” ๋ˆ„๊ตฌ๋‚˜ ๋ฐฐ์šธ ์ˆ˜ ์žˆ๊ณ 
11:41
that everybody can learn
183
701467
2043
๋ˆ„๊ตฌ๋‚˜ ์ž˜ ๋ฐฐ์šธ ์ˆ˜ ์žˆ๋‹ค๊ณ 
11:43
and learn well,
184
703534
1905
๋ฏฟ์—ˆ์Šต๋‹ˆ๋‹ค.
11:45
as long as we provide good learning materials
185
705463
3697
์šฐ๋ฆฌ๊ฐ€ ์˜จ๋ผ์ธ์œผ๋กœ ์ข‹์€ ๊ต์œก ์ž๋ฃŒ๋ฅผ
11:49
free on the web
186
709184
1455
๋ฌด๋ฃŒ๋กœ ์ œ๊ณตํ•œ๋‹ค๋ฉด ๊ฐ€๋Šฅํ•œ ์ผ์ž…๋‹ˆ๋‹ค.
11:50
so that they can access through the internet.
187
710663
3069
11:53
But such wonderful materials may benefit only those who can read well,
188
713756
5859
ํ•˜์ง€๋งŒ ํ›Œ๋ฅญํ•œ ์ž๋ฃŒ๋Š” ์ฝ๊ธฐ ์‹ค๋ ฅ์ด ์ข‹์€ ์‚ฌ๋žŒ๋“ค๋งŒ์ด ํ˜œํƒ์„ ๋ฐ›์„ ์ˆ˜ ์žˆ์ฃ .
12:00
and the percentage of those who can read well
189
720534
3935
์‚ฌ์‹ค ๊ทธ๋Ÿฐ ์‹ค๋ ฅ์„ ์ง€๋‹Œ ์‚ฌ๋žŒ์€
12:04
may be much less than we expected.
190
724493
3378
์šฐ๋ฆฌ๊ฐ€ ์˜ˆ์ƒํ•œ ์ˆ˜๋ณด๋‹ค ํ›จ์”ฌ ์ ์—ˆ์Šต๋‹ˆ๋‹ค.
12:10
How we humans will coexist with AI
191
730040
4241
์šฐ๋ฆฌ ์ธ๊ฐ„์ด AI์™€ ํ•จ๊ป˜ ์กด์žฌํ•œ๋‹ค๋Š” ๊ฒƒ์ด ๋ฌด์Šจ ์˜๋ฏธ์ธ์ง€
12:14
is something we have to think about carefully,
192
734305
3522
ํ™•์‹คํ•œ ์ฆ๊ฑฐ๋ฅผ ํ†ตํ•ด์„œ
12:17
based on solid evidence.
193
737851
2137
์‹ ์ค‘ํžˆ ์ƒ๊ฐํ•ด ๋ณด์•„์•ผ ํ•  ๊ฒƒ์ž…๋‹ˆ๋‹ค.
12:21
At the same time, we have to think in a hurry
194
741063
3977
๋™์‹œ์— ์šฐ๋ฆฌ๋Š” ์„œ๋‘๋ฅผ ํ•„์š”๊ฐ€ ์žˆ์Šต๋‹ˆ๋‹ค.
12:25
because time is running out.
195
745064
2402
์‹œ๊ฐ„์ด ์—†๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค.
12:28
Thank you.
196
748106
1162
๊ฐ์‚ฌํ•ฉ๋‹ˆ๋‹ค.
12:29
(Applause)
197
749626
3933
(๋ฐ•์ˆ˜)
12:34
Chris Anderson: Noriko, thank you.
198
754211
2080
ํฌ๋ฆฌ์Šค ์•ค๋”์Šจ(CA): ๋…ธ๋ฆฌ์ฝ”์”จ, ๊ฐ์‚ฌํ•ฉ๋‹ˆ๋‹ค.
12:36
Noriko Arai: Thank you.
199
756315
1765
๋…ธ๋ฆฌ์ฝ” ์•„๋ผ์ด(NA): ๊ฐ์‚ฌํ•ฉ๋‹ˆ๋‹ค.
12:38
CA: In your talk, you so beautifully give us a sense of how AIs think,
200
758104
5304
CA: ์ด๋ฒˆ ๊ฐ•์—ฐ์„ ํ†ตํ•ด AI๊ฐ€ ์–ด๋–ป๊ฒŒ ์ƒ๊ฐํ•˜๋Š”์ง€
12:43
what they can do amazingly
201
763432
1564
์–ผ๋งˆ๋‚˜ ๋†€๋ผ์šด ์ผ์„ ํ•˜๋Š”์ง€
12:45
and what they can't do.
202
765020
1695
๋ชปํ•˜๋Š” ๊ฒƒ์€ ๋ฌด์—‡์ธ์ง€ ์•Œ๊ฒŒ ๋˜์—ˆ์Šต๋‹ˆ๋‹ค.
12:46
But -- do I read you right,
203
766739
1494
ํ•˜์ง€๋งŒ ์ œ๊ฐ€ ์ดํ•ดํ•˜๊ธฐ์—
12:48
that you think we really need quite an urgent revolution in education
204
768257
5270
์šฐ๋ฆฌ์—๊ฒŒ ๊ต์œก ํ˜๋ช…์ด ์‹œ๊ธ‰ํ•œ ๊ณผ์ œ๋ผ๊ณ  ๋ง์”€ํ•˜์‹œ๋Š” ๊ฒƒ ๊ฐ™์€๋ฐ์š”.
12:53
to help kids do the things that humans can do better than AIs?
205
773551
4155
์•„์ด๋“ค์ด AI๋ณด๋‹ค ์ž˜ํ•  ์ˆ˜ ์žˆ๋Š” ๊ฒƒ์„ ์ฐพ๋„๋ก ๋„์™€์ฃผ์–ด์•ผ ํ•œ๋‹ค๋Š” ๋ง์”€์ด์ฃ ?
12:57
NA: Yes, yes, yes.
206
777730
1328
NA: ๋„ค, ๋„ค.
12:59
Because we humans can understand the meaning.
207
779082
4035
์šฐ๋ฆฌ ์ธ๊ฐ„์€ ์˜๋ฏธ๋ฅผ ์ดํ•ดํ•  ์ˆ˜ ์žˆ๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค.
13:03
That is something which is very, very lacking in AI.
208
783141
4906
AI์—๊ฒŒ๋Š” ์ „ํ˜€ ๋ถˆ๊ฐ€๋Šฅํ•œ ์ผ์ด์ฃ .
13:08
But most of the students just pack the knowledge
209
788071
4368
ํ•˜์ง€๋งŒ ๋Œ€๋ถ€๋ถ„์˜ ํ•™์ƒ ์—ญ์‹œ ์ง€์‹์„ ์ดํ•ดํ•˜๊ธฐ ๋ณด๋‹ค๋Š”
13:12
without understanding the meaning of the knowledge,
210
792463
3903
์ง€์‹์„ ๋‹จ์ง€ ํก์ˆ˜ํ•˜์ฃ .
13:16
so that is not knowledge, that is just memorizing,
211
796390
2809
๊ทธ๋ ‡๊ฒŒ ๋˜๋ฉด ์ง€์‹์ด ์•„๋‹Œ ๋‹จ์ˆœ ์•”๊ธฐ๊ฐ€ ๋˜๋Š” ๊ฒ๋‹ˆ๋‹ค.
13:19
and AI can do the same thing.
212
799223
2450
AI์™€ ๊ฐ™์€ ์ผ์„ ํ•˜๊ณ  ์žˆ๋Š” ์…ˆ์ด์ง€์š”.
13:21
So we have to think about a new type of education.
213
801697
3631
๋”ฐ๋ผ์„œ ์šฐ๋ฆฌ๋Š” ์ƒˆ๋กœ์šด ํ˜•ํƒœ์˜ ๊ต์œก์— ๋Œ€ํ•˜์—ฌ ์ƒ๊ฐํ•ด๋ด์•ผ ํ•ฉ๋‹ˆ๋‹ค.
13:25
CA: A shift from knowledge, rote knowledge, to meaning.
214
805352
3289
CA: ์•”๊ธฐ ์œ„์ฃผ์—์„œ ์˜๋ฏธ๋ฅผ ์ดํ•ดํ•˜๋Š” ๊ต์œก์œผ๋กœ ๋ณ€ํ™”์‹œํ‚ค๋Š” ๊ฒƒ์ด ์ค‘์š”ํ•˜๊ฒ ๋„ค์š”.
13:28
NA: Mm-hmm.
215
808665
1151
NA: ๋„ค. ๋งž์Šต๋‹ˆ๋‹ค.
13:29
CA: Well, there's a challenge for the educators. Thank you so much.
216
809840
3240
CA: ๊ต์œก์ž๋“ค์˜ ํฐ ์ˆ™์ œ๊ฒ ๊ตฐ์š”. ๋Œ€๋‹จํžˆ ๊ฐ์‚ฌํ•ฉ๋‹ˆ๋‹ค.
13:33
NA: Thank you very much. Thank you.
217
813104
1698
NA: ๊ฐ์‚ฌํ•ฉ๋‹ˆ๋‹ค.
13:34
(Applause)
218
814826
1185
(๋ฐ•์ˆ˜)
์ด ์›น์‚ฌ์ดํŠธ ์ •๋ณด

์ด ์‚ฌ์ดํŠธ๋Š” ์˜์–ด ํ•™์Šต์— ์œ ์šฉํ•œ YouTube ๋™์˜์ƒ์„ ์†Œ๊ฐœํ•ฉ๋‹ˆ๋‹ค. ์ „ ์„ธ๊ณ„ ์ตœ๊ณ ์˜ ์„ ์ƒ๋‹˜๋“ค์ด ๊ฐ€๋ฅด์น˜๋Š” ์˜์–ด ์ˆ˜์—…์„ ๋ณด๊ฒŒ ๋  ๊ฒƒ์ž…๋‹ˆ๋‹ค. ๊ฐ ๋™์˜์ƒ ํŽ˜์ด์ง€์— ํ‘œ์‹œ๋˜๋Š” ์˜์–ด ์ž๋ง‰์„ ๋”๋ธ” ํด๋ฆญํ•˜๋ฉด ๊ทธ๊ณณ์—์„œ ๋™์˜์ƒ์ด ์žฌ์ƒ๋ฉ๋‹ˆ๋‹ค. ๋น„๋””์˜ค ์žฌ์ƒ์— ๋งž์ถฐ ์ž๋ง‰์ด ์Šคํฌ๋กค๋ฉ๋‹ˆ๋‹ค. ์˜๊ฒฌ์ด๋‚˜ ์š”์ฒญ์ด ์žˆ๋Š” ๊ฒฝ์šฐ ์ด ๋ฌธ์˜ ์–‘์‹์„ ์‚ฌ์šฉํ•˜์—ฌ ๋ฌธ์˜ํ•˜์‹ญ์‹œ์˜ค.

https://forms.gle/WvT1wiN1qDtmnspy7